3-Benzoylquinoxalinone as a photoaffinity labelling derivative with fluorogenic properties allowing reaction monitoring under “no-wash” conditions

Article abstract of DOI:10.1039/d1cc01072g

Described herein is a quinoxalinone-based photoaffinity probe with caged fluorescence properties. Upon visible blue LED irradiation (λ_max450 nm), this photo-crosslinker is able to covalently capture proteins with concomitant fluorescence labelling. This process enables monitoring applications under “no wash” conditions.

Full text of DOI:10.1039/d1cc01072g

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
3-Benzoylquinoxalinone as a photoaffinity
labelling derivative with fluorogenic properties
allowing reaction monitoring under ‘‘no-wash’’
conditions†
Cite this: Chem. Commun., 2021,
57, 3893
Received 1st March 2021,
Accepted 12th March 2021
Madeleine Cauwel,a Clement Guillou,b Kevin Renault,a Damien Schapman,c
´
´
DOI: 10.1039/d1cc01072g
a
Magalie Benard,c Ludovic Galas,c Pascal Cosette,b Pierre-Yves Renard
and
´
a
rsc.li/chemcomm
Cyrille Sabot
*
Described herein is a quinoxalinone-based photoaffinity probe with in order to both increase the chance of the covalent bond
caged fluorescence properties. Upon visible blue LED irradiation formation once in close proximity and decrease the signal to
(kmax 450 nm), this photo-crosslinker is able to covalently capture noise ratio (if the electrophile reacts with other nucleophilic
proteins with concomitant fluorescence labelling. This process derivatives). On the other hand, in the affinity labelling (PAL)
enables monitoring applications under ‘‘no wash’’ conditions.
strategy, the electrophilic warhead is replaced by a more stable
photoreactive function, which unveiled highly reactive and tran-
The identification of protein–protein interactions or biological sient species only upon triggered UV irradiation. These inter-
targets of small-molecule ligands is of prime interest in medicinal mediates are able to react covalently with a larger range of
chemistry for elucidating new modes of action of pharmaco- residues, not limited to the nucleophilic amino acids. As a direct
phores. Important applications are also found in chemical biology consequence, this strategy has a broader application spectrum,
for the fundamental understanding of cellular functions or in while avoiding the biased reactivity of ABP towards specific
structural biology for identifying the amino acid residues of proteins equipped with the required nucleophilic moieties. The
proteins in contact with ligands.1
traditional photoreactive functions mostly consist of benzophe-
Methods for identifying biological targets include in vitro none, aryl azide, and diazirine groups that generate diradical,
activity assays, affinity chromatography,2 activity-based probe nitrene, and carbene intermediates, respectively, when exposed to
(ABP) technology,3 or photoaffinity labelling (PAL) approaches.4 280–365 nm UV irradiation. The access to functionalized chemical
Whilst the first two approaches rely on the use of isolated probes bearing a photoreactive function, a ligand, and a fluor-
proteins or crude cell lysates, in the latter two strategies, ophore is synthetically challenging, and the presence of a bulky
interactions from native environments could also be trapped fluorophore may disrupt protein–ligand interactions, and gener-
for their subsequent identification. On one hand, the activity- ally lead to low photo-crosslinking yields, which further compli-
based probes consist of (1) an affinity linker or ligand that cates the isolation and identification of the photo-crosslinked
favours interactions with the biological targets; (2) an electro- species. To address this point, minimalist probes were recently
philic warhead (such as an epoxide, a Michael acceptor, an designed, in which the fluorophore was elegantly replaced by a
activated halogen derivative, or a quinone methide) that reacts discrete bioorthogonal handle (i.e. terminal alkyne or cyclopro-
covalently and specifically with the nucleophilic residues of the pene) that was used to install the fluorescent tag once the
active site of the protein target; and (3) a reporter tag such as a photocrosslinking of the ligand–protein interactions was
fluorophore for the identification and purification of the modified achieved.5–7 However, this strategy requires a two-step protocol:
proteins. This strategy requires finding a fine tuning of the (1) the photolabelling step to ensure the integrity of the ligand–
respective reactivity of the two complementary reactive groups, protein complex during its isolation and (2) a bioorthogonal
reaction for installing the fluorescent tag. This approach can also
a Normandie Univ, CNRS, UNIROUEN, INSA Rouen, COBRA (UMR 6014),
Rouen 76000, France. E-mail: cyrille.sabot@univ-rouen.fr
b Normandie Univ, CNRS, UNIROUEN, INSA Rouen, PBS-UMR6270 CNRS,
PISSARO Proteomics Facility, IRIB, Rouen 76000, France
c Normandie Univ, Inserm, UNIROUEN, PRIMACEN, Cell Imaging Platform of
Normandy, IRIB, Rouen 76000, France
produce a background fluorescence due to the presence of the
residual fluorescent tag typically used in excess, which prevents,
among others, real-time monitoring applications. An alternative,
yet underexplored approach consists in using fluorogenic photo-
reactive functions, which are able to both photocrosslink and
unveil a fluorescent scaffold in a single-step process. This
approach has been combined with diazo-,8 azide-8–11 and
† Electronic supplementary information (ESI) available. See DOI: 10.1039/
d1cc01072g
This journal is © The Royal Society of Chemistry 2021
Chem. Commun., 2021, 57, 3893–3896 | 3893

View Article Online
Communication
ChemComm
Fig. 1 Fluorogenic photoreactive functions.
diazirine-based photoreactive functions.12,13 This strategy was
extended to the photoreactive diaryltetrazole function, which has
been shown to preferentially react with aspartic acid and glutamic
acid residues of proteins through the formation of a nitrile imine
intermediate.14 However, 280 and 300 nm UV irradiation are
required to photoactivate aryl azides and diaryltetrazoles, respec-
tively, which can induce severe damage to biological systems (Fig. 1).
Among the usual photo-crosslinkers, the benzophenone class
of compounds is scarcely represented yet; only
a photo-
crosslinking system with uncaged fluorescence properties has
been reported. In this system, the aryl ketone group was incorpo-
rated into a bulky BODIPY fluorophore.15 A small fluorogenic
diaryl ketone-based photoreactive function would be, however,
particularly useful for the direct and real-time monitoring of
photo-affinity crosslinking through the concomitant fluorescent
labelling experiments. In fact, the aryl azide and diazirine deriva-
tives generate irreversibly reactive intermediates upon UV activa-
Fig. 2 (A) Preparation of 3-benzoylquinoxalinones 1 and 3 and hydroxyl-
containing quinoxalinones 2 and 4. Reagents and conditions: (a) NaBH4
and MeOH, room temperature for 10 min (98%, 2; quant., 4). (B) Photo-
reaction of quinoxalinone 5 with pyrrolidinone under visible blue LED light
(lmax 450 nm) for 2 h; (C) RP-HPLC fluorescence chromatograms obtained
before and after irradiation of 5 in pyrrolidinone (lex 375 nm and lem 450 nm).
tion that react with the surrounding water molecules in the the formation and troublesome separation of isomeric inter-
absence of nearby biological targets, leading to fluorescence mediates. Compound 3 was prepared in a total of 3 steps in
recovery.12 On the other hand, once photo-activated, benzophe- 29% overall yield from readily available 4,5-dimethoxy-1,
none derivatives yield diradical intermediates, which either react 2-phenylenediamine. The synthetic strategy relies on the ring
with the desired target or react reversibly with water molecules, contraction of benzodiazepine-2-ones that we previously des-
thus preventing the formation of a fluorescent product in the cribed mediated by N-bromosuccinimide in the presence of
absence of a nearby protein target.16
DMSO, which serves both as a solvent and an oxidant.19
Then, the carbonyl to hydroxyl fluorescence turn-on ability
Herein, we report the use of a benzophenone-quinoxalinone
hybrid system for the fluorogenic photoaffinity labelling of of 3-benzoylquinoxalinone 3 functionalized by two electron-
protein targets. As an additional bonus, the labelling reaction donating methoxy substituents was investigated. As expected,
could proceed above 400 nm, conditions that are less detri- the treatment of 3 by NaBH4 resulted in a remarkable 100-fold
mental for the protein structures.17
fluorescence enhancement. This was also consistent with the
We recently reported that 3-benzoyl-6-methoxyquinoxalin-2(1H)- values of the fluorescence quantum yields of 3 and 4, which
ones such as compound 1 exhibited a dramatic fluorescence were determined to be o0.01 and 0.35 in PBS pH 7.4, respec-
turn-on upon NaBH4 reduction of the ketone moiety into the tively. Importantly, compound 3 possesses a large absorption
corresponding alcohol (Fig. 2A).18 The presence of the ketone band with a maximum at 402 nm, which will allow the irradia-
moiety is responsible for the fluorescence quenching, presumably tion of quinoxalinone probes in the visible region (Fig. S4,
through a singlet-to-triplet intersystem crossing process, a non- ESI†), thus limiting the photodamage induced by o350 nm
radiative de-excitation pathway which competes with the fluores- irradiations. To investigate the benzophenone-like behaviour of
cence phenomenon.18 From these studies, it is therefore expected such benzoylquinoxalinones upon irradiation, a similarly pre-
that the transformation of the carbonyl group during photoaffinity pared compound 5 (bearing an ester moiety useful for the
labelling reactions would also lead, to some extent, to fluorescence future functionalization) was photoexcited by a visible blue
recovery. However, the access to compound 1 required the separa- LED light (lmax 450 nm) in neat pyrrolidinone (or in a mixture
tion of the isomeric 7-methoxy- and 8-methoxy-benzodiazepinone with water), which was used as a simple amide-containing the
intermediates at the N-methylation step. Accordingly, we decided to peptide linkage. The expected crosslinked product 6 was not formed,
prepare the 6,7-dimethoxyquinoxalinone analogue 3, thus avoiding but instead compound 7 lacking the benzoyl group was obtained as
This journal is © The Royal Society of Chemistry 2021
3894 | Chem. Commun., 2021, 57, 3893–3896

View Article Online
ChemComm
Communication
the only fluorescent product (Fig. 2B and C). Two plausible
mechanisms were proposed involving either a Norrish Type I
photochemical cleavage of 5 or the formation of a diradical of
imine at the first step (Fig. S33 and S34, ESI† for mechanistic
details).
Following this preliminary study, the fluorescent labelling of
bovine carbonic anhydrase II (CA-II, molecular weight 30 kDa)
used as a model target protein was investigated. From quinox-
alinone 5, a two-step ester hydrolysis/amidation sequence was
carried out to introduce the 4-sulfamoyl benzylamine ligand,
which belongs to a well-established family of CA-II inhibitors,
leading to probe 8.20 In order to investigate the influence of the
chain length on the labelling efficiency, probe 9, in which the
ligand is separated by two PEG units from the photoreactive
quinoxalinone scaffold, was also accordingly prepared
(Scheme 1, see ESI† for experimental details).
With these compounds in hands, their IC50 values were
determined and found to be 1.19 mM and 1.46 mM for 8 and
9, respectively. These values are in agreement with the one
found for the parent compound benzylamide 10 (IC50 = 1.90 mM),
Fig. 3 (A) Fluorescent emission spectra of
a
mixture of CA-II
(10 mM ) with different probes (10 mM, 1 equiv.)
=
0.3 mg mLÀ1
thus demonstrating that the 3-benzoylquinoxalinone scaffold (lex 375 nm) in water at 25 1C, before (dashed line) and after (solid line)
1 h of blue LED light irradiation: (a) probe 8; (b) probe 9; (c) quinoxalinone
3; and (d) 30 min pre-incubation with known CA-II inhibitor acetazolamide
does not significantly interfere with the binding properties of
10, whatever the chain length. The photo-crosslinking and con-
(100 mM, 10 equiv.) before introducing probe 8. (B) SDS-PAGE analysis of
comitant fluorescent labelling of bovine CA-II (10 mM) with probes
8 and 9 was conducted in PBS pH 7.4 under visible blue LED
irradiation (lmax 450 nm). The photoaffinity labelling could
be followed by fluorescence spectroscopy (lex 375 nm) under
‘‘no-wash’’ conditions, that is, without filtering the excess of
CA-II (30 kDa). Conditions: 8 (10 mM), 3 (10 mM), CA-II (10 mM = 0.3 mg mLÀ1
)
in water and LED blue light irradiation (1 h). *30 min pre-incubation with a
known CA-II inhibitor acetazolamide (100 mM, 10 equiv.) before introducing
probe 8 (10 mM). In-gel fluorescence imaging (lex = 302 nm and lem 4535 nm).
unreacted probe. In fact, no fluorescence increase was observed Of note, two emission peaks at 460 and 560 nm were observed
when 3-benzoyquinoxalinone is irradiated for 1 h in PBS pH 7.4 for quinoxaline derivatives in aqueous systems. The comple-
(Fig. S12, ESI†). In contrast, as shown in Fig. 3A-a and b, the mentary photophysical studies suggest that one of the two
probes 8 and 9 showed a noticeable 10- and 3-fold increase, peaks may correspond to that of aggregates of the quinoxali-
respectively, in fluorescence at 460 nm upon binding to the none molecules. In fact, only one fluorescence emission band
protein. The labelling efficiency with probe 8 was found to be was observed in the organic solvent DMSO, in which quinox-
about 2% (Fig. S19, ESI†). Most of the labelling was achieved after alinones are well soluble (Fig. S8, ESI†).
15 min and 3 min of irradiation with visible blue LED
(lmax 450 nm) and under near UV LED (lmax 405 nm), respectively further the fluorescent labelling of CA-II specifically with 8
(Fig. S14–S16, ESI†).
and under light irradiation stimulus, with the presence of a
The control experiments carried out with 3 containing only fluorescent band in the 30 kDa region (Fig. 3B, lane 2).
the photoreactive part of the probe (Fig. 3A-c), or with the probe
The labelling specificity was further investigated by tagging
Moreover, the gel electrophoresis analyses confirmed
8 in the presence of CA-II inhibitor, acetazolamide (10 equiv., the bovine CA-II mixed with human blood plasma (HBP, Fig. 4).
Fig. 3A-d), showed, however, no significant fluorescence Similarly, CA-II was efficiently fluorescently labelled by probe 8,
increase. These results indicate the specificity of the labelling. while only light fluorescence for CA-II was detected with the
non-specific probe 3. In the absence of irradiation, no labelling
was observed with 8, confirming that this probe provides
spatial and temporal control over reactivity.
Encouraged by these results, the fluorescence turn-on properties
of quinoxalinones were next investigated in a more complex
medium, through the fluorescent labelling of cells under ‘‘no-
wash’’ conditions. To illustrate that proof-of-principle, the func-
tionalizable quinoxalinone 5 was equipped with a phospholipid,
1,2-dioleoyl-sn-glycero-3-phosphoethanolamine known to interact
with the biological membranes, yielding a phospholipid-based
probe 11, in order to specifically visualize the lipid-rich organelles
that are abundant in cells (Fig. 5).21,22 After incubation of fixed
Scheme 1 Preparation of the modified quinoxalinones 8 and 9.
cells with probe 11 and visible blue LED irradiation, both for 1 h,
This journal is © The Royal Society of Chemistry 2021
Chem. Commun., 2021, 57, 3893–3896 | 3895

View Article Online
Communication
ChemComm
LED, conditions that are less photo-damaging to proteins and
cell contents than traditionally used UV irradiations. The use of
the carbonyl-conjugated quinoxalinone systems possessing photo-
crosslinking ability accompanied by the caged-fluorescence pro-
perties enables the direct monitoring of the fluorescent labelling,
that is, without removing the excess of the unreacted probe. This
was leveraged for the fluorescent labelling of the endomembrane
systems of cells by using a photoreactive phospholipid probe. This
proof-of-concept being established, the 3-benzoylquinoxalinone-
based probes will be used for identifying new protein–protein
interactions occurring in specific subcellular organelles, and the
results will be published in due course.
Fig. 4 Specific fluorescent labelling of CA-II (30 kDa) by 8 in the presence
of human blood plasma. Conditions: 3 (100 mM) or 8 (100 mM), CA-II
This work was supported by the Region Normandie, and PhD
(100 mM = 3 mg mLÀ1), human blood plasma diluted by 10 in PBS pH 7.4 and financial support to MC. This work was also partially supported by
visible blue LED light irradiation (2 h). In-gel fluorescence imaging (lex = 302 nm,
the Centre National de la Recherche Scientifique (CNRS), the INSA
lem 4 535 nm).
Rouen, Rouen Normandy University, The Institute of Research
and Innovation in Biomedicine (IRIB), and the Labex SynOrg
(ANR-11-LABX-0029). We also thank Laetitia Bailly (INSA Rouen)
and Emilie Petit (INSA Rouen) for the HPLC-MS analyses and
Albert Marcual (CNRS) for the HRMS analyses.
Conflicts of interest
There are no conflicts to declare.
Notes and references
1 E. L. Vodovozova, Biochemistry, 2007, 72, 1–20.
2 D. V. Titov and J. O. Liu, Bioorg. Med. Chem., 2012, 20, 1902–1909.
3 E. Smith and I. Collins, Future Med. Chem., 2015, 7, 159–183.
4 H. Guo and Z. Li, MedChemComm, 2017, 8, 1585–1591.
5 Z. Li, P. Hao, L. Li, C. Y. Tan, X. Cheng, G. Y. Chen, S. K. Sze,
H. M. Shen and S. Q. Yao, Angew. Chem., Int. Ed., 2013, 52, 8551–8556.
6 Z. Li, D. Wang, L. Li, S. Pan, Z. Na, C. Y. Tan and S. Q. Yao, J. Am.
Chem. Soc., 2014, 136, 9990–9998.
7 S. Kellner, S. Seidu-Larry, J. Burhenne, Y. Motorin and M. Helm,
Nucleic Acids Res., 2011, 39, 7348–7360.
8 S.-Y. Dai and D. Yang, J. Am. Chem. Soc., 2020, 40, 17156–17166.
ˇ
9 E. Paunescu, L. Louise, L. Jean, A. Romieu and P.-Y. Renard,
Dyes Pigm., 2011, 91, 427–434.
10 K. Chiba, M. Asanuma, M. Ishikawa, Y. Hashimoto, K. Dodo,
M. Sodeoka and T. Yamaguchi, Chem. Commun., 2017, 53, 8751–8754.
11 K. Chiba, Y. Hashimoto and T. Yamaguchi, Chem. Pharm. Bull., 2017, 65,
994–996.
12 T. Tomohiro, A. Yamamoto, Y. Tatsumi and Y. Hatanaka,
Chem. Commun., 2013, 49, 11551–11553.
Fig. 5 (A) Labelling cells with irradiated probe 11 (10À3 M). Scale bar = 50 mm
(B) influence of probe concentration. ***, p o 0.001, statistical significance
versus probe 11 without irradiation; *, p o 0.05, statistical significance versus
probe 11 without irradiation (C and D). The confocal images at different
z position (z1 and z2). Scale bar = 10 mm (C) and scale bar = 5 mm (D). The
white arrows indicate a labelled membrane of the intracellular vesicles.
13 T. M. Ayele, S. D. Knutson, S. Ellipilli, H. Hwang and J. M. Heemstra,
Bioconjugate Chem., 2019, 30, 1309–1313.
14 Z. Li, L. Qian, L. Li, J. C. Bernhammer, H. V. Huynh, J. S. Lee and
S. Q. Yao, Angew. Chem., Int. Ed., 2016, 55, 2002–2006.
15 D. P. Murale, S. C. Hong, J. Yun, C. N. Yoon and J. S. Lee,
Chem. Commun., 2015, 51, 6643–6646.
16 G. Dorman, H. Nakamura, A. Pulsipher and G. D. Prestwich,
Chem. Rev., 2016, 116, 15284–15398.
17 D. I. Pattison, A. S. Rahmanto and M. J. Davies, Photochem. Photo-
biol. Sci., 2012, 11, 38–53.
18 K. Renault, P.-Y. Renard and C. Sabot, New J. Chem., 2017, 41,
10432–10437.
19 H. Mtiraoui, K. Renault, M. Sanselme, M. Msaddek, P. Y. Renard
and C. Sabot, Org. Biomol. Chem., 2017, 15, 3060–3068.
20 C. B. Mishra, S. Kumari, A. Angeli, S. M. Monti, M. Buonanno,
M. Tiwari and C. T. Supuran, J. Med. Chem., 2017, 60, 2456–2469.
21 J. Yang, J. Seckute, C. M. Cole and N. K. Devaraj, Angew. Chem., Int.
Ed., 2012, 51, 7476–7479.
confocal microscopy showed a concentration-dependent effective
labelling of most, if not all, PC12 cells (Fig. 5). Satisfyingly, a
significantly lower fluorescence was observed in the control
experiment carried out in the absence of irradiation (Fig. 5B). In
addition, higher magnifications revealed subcellular positive ele-
ments, suggesting that the endomembranes around the nuclear
membrane as well as the membranes of the intracellular vesicles
are labelled with probe 11.
In summary, an easy photolabelling procedure in the presence
of compact quinoxalinone derivatives playing the role of both the
photoreactive function and pro-fluorophore, is reported. Advanta-
geously, the labelling proceeds effectively under a visible blue
22 X. Ji, K. Ji, V. Chittavong, R. E. Aghoghovbia, M. Zhu and B. Wang,
J. Org. Chem., 2017, 82, 1471–1476.
This journal is © The Royal Society of Chemistry 2021
3896 | Chem. Commun., 2021, 57, 3893–3896